The hypothesis of this protocol is that increases in HIV-1 viral load in the central nervous system (CNS) of infants and children infected with the virus will be associated with a poor neurological outcome. Also, we will examine that as more children receive combination therapy with antiretrovirals that have varying degrees of CNS penetration, the CNS will act as a reservoir for virus and viral RNA from the CNS will have different resistance patterns than isolates from the peripheral blood. We also propose that virus isolated from the CSF of children with HIV encephalopathy is genetically different from virus in the CSF of neurologically asymptomatic children.